考虑个体车辆先验数据的交通信号模糊控制研究  

Research on Traffic Signal Fuzzy Control Considering Individual Vehicle Prior Data

作　　者：章锡俏[1] 崔乐祺 杜佳明 赵江 ZHANG Xi-qiao;CUI Le-qi;DU Jia-ming;ZHAO Jiang(School of Transportation Science and Engineering,Harbin Institute of Technology,Harbin 150090,Heilongjiang,China)

机构地区：[1]哈尔滨工业大学交通科学与工程学院,黑龙江哈尔滨150090

出　　处：《中国公路学报》2023年第10期305-316,共12页China Journal of Highway and Transport

基　　金：国家重点研发计划项目(2020YFB1600400);黑龙江省自然科学基金项目(LH2019E052)。

摘　　要：提出一种基于车辆行程时间和路径预测的交通信号模糊控制方法,以充分利用个体车辆的历史出行信息进行信号控制优化。首先,利用贝叶斯优化的支持向量机模型,结合车辆个体特征和外部环境特征,预测个体车辆从上游到达目标交叉口的行程时间。其次,基于马尔可夫链建立贝叶斯车辆路径预测模型,根据实时车牌识别结果预测车辆出行路径,分析车辆转向选择行为。然后,根据个体车辆到达交叉口的时间和转向,集计得到交叉口各进口道各方向的预测到达率,模糊控制器根据预测到达率,输出绿灯延时,决定下一执行相位。最后,通过SUMO仿真平台对基于车辆行程时间和路径预测的交通信号模糊控制方法进行验证与效益分析。分析中,将所提出的方法与传统固定到达率的模糊控制方法进行比较,分别对平峰时段和高峰时段进行案例研究。研究结果表明,相较于传统模糊控制方法,所提出的新方法的平峰时刻车均延误下降了18.75%,高峰时刻车均延误下降了16.11%,同时,灵活相序的控制方式提高了绿灯时间的利用效率。This study proposes a traffic signal fuzzy control method based on vehicle travel time and route prediction to fully utilize the historical travel information of individual vehicles for signal control optimization.First,the travel time of individual vehicles from upstream to the target intersection was predicted using the Bayesian optimized support vector machine model combined with individual vehicles and external environmental characteristics.Second,a Bayesian vehicle route prediction model was established based on the Markov chain,predicting vehicle travel routes and analyzing vehicle turning behaviors according to real-time license plate recognition results.Subsequently,based on the time and turning of individual vehicles arriving at the intersection,the predicted arrival rate in each direction at each intersection entrance was aggregated,and the fuzzy controller outputted the green light delay based on the predicted arrival rate,determining the next execution phase.Finally,the traffic signal fuzzy control method based on vehicle travel time and route prediction was verified,and a benefit analysis was conducted using the SUMO simulation platform.In the analysis,the proposed method was compared with the traditional fixed arrival rate fuzzy control method using case studies conducted during both off-peak and peak periods.The research results show that the new method proposed in this study,compared with the traditional fuzzy control method,reduces the average vehicle delay by 18.75%during off-peak times and 16.11%during peak times,whereas the flexible phase sequence control method improves the utilization efficiency of the green light time.

关 键 词：交通工程 交通信号控制 模糊控制 个体车辆数据 行程时间预测 路径预测 

分 类 号：U491[交通运输工程—交通运输规划与管理]